Different CEOs have different ideas on how AI will impact the workplace.

Bhavish Aggarwal, CEO of Ola Electric and founder of AI startup Krutrim, spent last week building AI agents for Ola Electric — and came away with a pointed observation about corporate structure. “So many layers get built between the actual doers and the founder as the company scales,” he wrote on X. His conclusion: AI agents will eliminate the middlemen — specifically, the ones whose primary function is managing people rather than solving problems.

“Agents will take away all middlemen in a company who are only ‘managing people’ and not doing any problem solving,” Aggarwal wrote. “And the people actually building will be even more valuable.”

It’s a view that’s gaining traction among technology leaders, though the framing varies. NVIDIA CEO Jensen Huang has said that IT departments of the future will essentially become HR departments for AI agents — managing, onboarding, and nurturing digital workers rather than human ones. Upwork CEO Hayden Brown made a related argument when laying off 25% of his workforce earlier this year, writing that AI makes it possible for smaller teams to have a bigger impact than ever. Jack Dorsey, in announcing Block’s own sweeping cuts, said AI now does the information-routing work that layers of management once existed to perform.

What’s notable about Aggarwal’s post is that he’s arriving at these conclusions from hands-on use, not theory. He described jumping on the “vibe coding bandwagon” and building a bunch of AI agents himself this past week — a founder personally road-testing the tools, then drawing structural conclusions from what he built.

Aggarwal has been one of India’s more vocal advocates for AI adoption. His company Krutrim — now India’s first AI unicorn — was founded on the premise that India needs its own AI infrastructure, built for Indian languages and use cases. Whether Krutrim has the technical depth to compete with global labs remains an open question; Aggarwal himself acknowledged last year that the company was “nowhere close to global benchmarks yet”. But his broader argument about AI reorganizing companies from the inside out is one that a growing number of executives are making in concrete terms.

Citadel CEO Ken Griffin recently described watching PhD-level finance work — tasks that would have taken months of human effort — being completed by AI agents in days. Goldman Sachs CEO David Solomon has pushed back on the more alarmist takes, arguing that when people are freed from routine tasks, they tend to take on more complex, higher-value work rather than disappear from the workforce.

Where Aggarwal’s view differs from Solomon’s is in who he thinks is at risk. He’s not talking about AI replacing frontline workers or technical contributors. His argument is more targeted: the people who exist primarily to manage other people, without doing substantive problem-solving themselves, are the ones whose roles AI will absorb. The people who are actually building things, in his view, become more valuable as a result.

That’s a thesis plenty of founders and operators would quietly agree with, even if they wouldn’t say it publicly.